"TypeError: Argument 'thresh' must be double, not bool" on Recognition Using CTC Model

[Roy Ritche Vy]

I am trying to run the model using the OCR notebook and I was able to run the code smoothly until the last cell. Its giving me the error "TypeError: Argument 'thresh' must be double, not bool" pointing to the last line of code. Am I missing something here?

[ori30ffs]

pip install opencv-python==4.2.0.34 helped me

[Cesar M.]

there are more sofisticated way.
Need change some lines, that i did:

• return _crop_add_border(th, height, border, border_size)

• return _crop_add_border(th, height=height, border=border, border_size=border_size)

The function in question is "word_normalization", int "./src/ocr/normalization.py"

It will be like this:

def word_normalization(image, height, border=True, tilt=True, border_size=15, hyst_norm=False):
    """ Preprocess a word - resize, binarize, tilt world."""
    image = resize(image, height, True)
    
    if hyst_norm:
        th = _hyst_word_norm(image)
    else:
        img = cv2.bilateralFilter(image, 10, 30, 30)
        gray = 255 - cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
        norm = cv2.normalize(gray, None, 0, 255, cv2.NORM_MINMAX)
        ret,th = cv2.threshold(norm, 50, 255, cv2.THRESH_TOZERO)

    if tilt:
        return _word_tilt(th, height, border, border_size)
    return _crop_add_border(th, height=height, border=border, border_size=border_size)

Hope this helps